to produce the most relevant search results; otherwise, users would switch to their competitors. 17 However, markets alone are unlikely to address the concerns of targeted manipulation for three reasons:

Proprietary algorithms. Users would not be able to detect targeted manipulation in most cases, as search engines keep their algorithms secret. 3 Moreover, even if users are more aware of results manipulation, their expectation of what they are searching for is shaped during the search process. 17 Consequently, they cannot objectively evaluate search-engine quality17;

Concentrated market. While the Internet search market is highly concentrated, a less monopolistic market is unlikely to emerge due to high economies of scale. 3 Furthermore, incumbents benefit from their existing user base by, say, collecting user data through such products as the Google toolbar. Moreover, the emergence of new big players is also unlikely, since promising start-ups could be acquired by such dominant search giants as Google and Microsoft; and

User inertia. While switching might seem easy (users simply type another address), evidence suggests that personal habit is a key factor when selecting a search engine22; moreover, new technologies like personalized search are likely to raise switching costs significantly.

Technological development is also often mentioned in the context of search-engine bias. 12 However, while new technologies may alleviate some concern (such as reinforcement of popular sites), no technology in sight is likely to cure the problem of targeted manipulation.

Two counterarguments often brought up against regulation of search-engine bias are that search results are free speech and therefore cannot be regulated, and search engines are not essential facilities, as they do not fulfil the criteria of essential facilities accepted by U.S. courts.

While these arguments have merit, they are insufficient for rejecting intervention for several reasons: Though the courts have acknowledged First Amendment rights for search engines, a number of legal scholars have argued against this view. 3, 5 Both

markets alone
are unlikely to be
sufficient to address
the concerns
about targeted
manipulation.

arguments emerge from a U.S.-centric context, which, as the Microsoft antitrust case in the 1990s showed, is not the only legal arena for regulation. Moreover, in legal circles it has been suggested that the existing regulatory frameworks may be inadequate for something as groundbreaking as Internet search. Given the state of the law, governments and multinational bodies may need to create a new regulatory framework.

search and its stakeholders

Who are the stakeholders and what are their interests? Mitchell wrote19 that stakeholders are characterized by power, legitimacy, and urgency. Building on a broad investigation of stakeholder interests in search-engine law by Grimmelmann, 15 we see four main actors—users, search engines, Webmasters, and search-engine optimizers—that are, to some extent, characterised by Mitchell’s attributes.

Table 1 is an overview of key stakeholder interests in Internet search bias and stakeholder recognition of possible manipulation. Search-engine optimizers are particularly conflicted. On the one hand, they stand to gain from greater transparency in Internet search, as their business would be easier and more efficient, and a clearer picture of accepted practices would enable them to guarantee their clients that search engines tolerate their techniques. On the other hand, they profit from lack of transparency, as it raises the value of their key assets—expertise and inside knowledge. Moreover, search-engine optimizers with good contacts and community-wide attention can profit from their influence with search-engine companies in “curing” cases of rank demotion. Therefore, top performers arguably have little interest in greater transparency.

Commentators have made several proposals for regulating targeted manipulations. We comment first on the two most promising, then introduce a new approach (outlined in Table 2).

Obligation to provide reasons for rank demotion. One proposal20 suggested establishing an obligation to provide a reason for rank demotion to increase transparency in the relationship between Webmasters and search